What the AI found
“Your recovery heart rate has been, on average, 9 BPM higher on days following evening interval sessions after 7 PM, compared to earlier training times.”
Before
Ad Hoc Training Log & Vague Feelings
After
Structured Weekly Recovery Insights
The same system, three states — real screens, not a screenshot
| 18 Jan | RHR 52, HRV 45. Felt tired, hard interval session. |
| 19 Jan | RHR 55, HRV 38. Sleep score 68. Evening gym. |
| 20 Jan | RHR 51, HRV 50. Easy run, good sleep. |
| 21 Jan | RHR 53, HRV 42. Long bike, felt flat. |
Prompt
Here is four weeks of my daily recovery data, including Resting Heart Rate (RHR), Heart Rate Variability (HRV), sleep score, and notes on training type and time. Identify any strong correlations between training time (specifically evening sessions after 7 PM vs. earlier) and my recovery metrics, especially RHR and HRV. Provide specific patterns or averages.
Here is four weeks of my daily recovery data, including Resting Heart Rate (RHR), Heart Rate Variability (HRV), sleep score, and notes on training type and time. Identify any strong correlations between training time (specifically evening sessions after 7 PM vs. earlier) and my recovery metrics, especially RHR and HRV. Provide specific patterns or averages.
AI
Analysis of your last four weeks indicates a notable pattern: your average Resting Heart Rate (RHR) on days following evening interval or threshold sessions (after 7 PM) was 9 BPM higher (average 56 BPM) compared to days following training sessions completed before 5 PM (average 47 BPM). Heart Rate Variability (HRV) showed a corresponding, albeit less pronounced, decrease of approximately 7 ms (average 38 ms vs. 45 ms) after late evening intense efforts. This suggests a potential link between late-night intensity and compromised recovery.56 BPM
Avg. RHR (Late Training)
47 BPM
Avg. RHR (Early Training)
38 ms
Avg. HRV (Late Training)
Weekly Recovery Check-ins: From Scattered to Structured
A triathlete moves from disjointed recovery notes to a structured weekly review, spotting a crucial correlation between late-night training and recovery heart rate.
A 38-year-old amateur triathlete, Northern Europe
Tools used
The real tools used here — swap any for your own equivalent. Each links to how we’d set it up.
Starting state
Before anything was set up
Our triathlete maintained a diligent, if somewhat scattered, training log across a few apps. Daily notes on perceived exertion, sleep quality, and general mood were present, but the sheer volume of unstructured data made it difficult to connect the dots. Reviewing meant scrolling through endless entries, hoping a pattern would emerge through sheer willpower. It was clear there were trends, but extracting them felt like guesswork.
| 18 Jan | RHR 52, HRV 45. Felt tired, hard interval session. |
| 19 Jan | RHR 55, HRV 38. Sleep score 68. Evening gym. |
| 20 Jan | RHR 51, HRV 50. Easy run, good sleep. |
| 21 Jan | RHR 53, HRV 42. Long bike, felt flat. |
| 22 Jan | RHR 56, HRV 35. Evening threshold run. Poor sleep. |
Working state
Membership, doing its job
Seeking a more systematic approach, our triathlete copied their last four weeks of daily recovery metrics into a Google Sheet and turned to Gemini. The goal was to consolidate qualitative observations with quantitative data from their Oura ring and Garmin watch. By providing a clear prompt, the AI began to sift through the daily entries, looking for non-obvious correlations that would otherwise be buried in the noise.
Prompt
Here is four weeks of my daily recovery data, including Resting Heart Rate (RHR), Heart Rate Variability (HRV), sleep score, and notes on training type and time. Identify any strong correlations between training time (specifically evening sessions after 7 PM vs. earlier) and my recovery metrics, especially RHR and HRV. Provide specific patterns or averages.
Here is four weeks of my daily recovery data, including Resting Heart Rate (RHR), Heart Rate Variability (HRV), sleep score, and notes on training type and time. Identify any strong correlations between training time (specifically evening sessions after 7 PM vs. earlier) and my recovery metrics, especially RHR and HRV. Provide specific patterns or averages.
AI
Analysis of your last four weeks indicates a notable pattern: your average Resting Heart Rate (RHR) on days following evening interval or threshold sessions (after 7 PM) was 9 BPM higher (average 56 BPM) compared to days following training sessions completed before 5 PM (average 47 BPM). Heart Rate Variability (HRV) showed a corresponding, albeit less pronounced, decrease of approximately 7 ms (average 38 ms vs. 45 ms) after late evening intense efforts. This suggests a potential link between late-night intensity and compromised recovery.Use case implemented
The finished system, running on its own
Now, with a clear summary available each Sunday, the triathlete spends just minutes pinpointing trends. The focus has shifted from manual data hunting to acting on clear insights. This structured feedback loop provides actionable intelligence, allowing for informed adjustments to their training schedule and recovery protocols, and turning subjective feelings into objective data points for better decision-making.
56 BPM
Avg. RHR (Late Training)
47 BPM
Avg. RHR (Early Training)
38 ms
Avg. HRV (Late Training)
What an outside observer would notice
60 mins to 8 mins
Weekly review time
-5 BPM
Adjusted RHR after late-night training
3 per week
Training session re-scheduling
The stack — build it yourself
Familiar, free, and excellent for consolidating data from various sources into a single, structured format for AI analysis.
Its natural language understanding allows complex queries against raw data, surfacing non-obvious correlations that would be time-consuming to find manually.
Provides passive, reliable data on RHR, HRV, and sleep, forming essential inputs for recovery analysis.
Automates the tracking of training load, type, and timing, offering crucial context for recovery data.
These are the tools used in this story. Any can be swapped for an equivalent you already trust.
Go deeper
Do this yourself
Read the full story: From Scattered Notes to Structured Insights
This story runs on Membership. The tools and prompts above are the real build — swap any tool for your own equivalent and follow the same steps.